Aristotle's "Two" Views of Induction: McCaskey's Resolution (Part 1)

In my studies of Objectivism and Aristotle, I’ve always thought that I understood what I was reading about and contemplating afterward. One of the exceptions to that understanding was my extended reading of Aristotle’s concept of “induction.”

My earlier note, “Aristotle on Induction,” was really one interpretation of his position, supported by the vast majority of Aristotle’s use of the term “induction” (epagōgē) in his known works. What I didn’t include in that note was the alternative view, the contemporary and dominant view, defended chiefly by citing Prior Analytics, book 2, chapter 23 (PrA 2.23); an account which seemingly presents an entirely different, even contradictory, view of induction than what is presented in Aristotle’s other works (and even the understanding of induction presented within the first paragraph of that 23rd chapter ). Upon reading this chapter, I wrote in my MS Word document for Aristotle notes: “I have problems with the view that Aristotle thought induction required ‘complete enumeration of instances’; it only definitely shows up twice, and isn’t consistent with his other statements on induction.” I simply could not connect this position on induction with, for instance, the views of induction presented in his Topics or Rhetoric.

Philosopher John McCaskey, in the first chapter of his unpublished dissertation “Regula Socratis: The Rediscovery of Ancient Induction in Early Modern England,” attempts to reconcile these two conflicting views by arguing that one position is actually Aristotle‘s position, and that the other position is simply a misunderstanding and misreading of Aristotle.

Before discussing the resolution, let’s review the details of these differing views on induction.

The Two Interpretations

McCaskey notes that the term “epagōgē” (induction) appears 96 times in Aristotle’s known works. The first interpretation, the one McCaskey accepts (along with a very few others in the history of philosophy, including myself), can be understood by reading the majority of those 96 uses of “induction.” It is primarily presented in the Topics, Rhetoric, and Posterior Analytics 2.19. This interpretation is summarized as follows:

Induction is a form of reasoning that moves from particular instances and rises to general and universal knowledge. It is founded on sense-perception and memory: from our senses and memories we gain experience, and from these experiences we gain universal knowledge via induction. The application of an induction extends beyond the particulars that were used in its formation, meaning that induction is an open-ended process, rather than limited to the two or three particulars that were used to form the inductive generalization. It is different from deduction, and a counterpart to that other form of reasoning. More precisely, induction is the fundamental method of reasoning in comparison to deduction, because it supplies the premises for deductions. Induction gets its force or legitimacy from the similarity of particulars, not their number. Lastly, induction is a tool for making conceptual generalizations by identifying the essential nature of things (this last from McCaskey, page 35 of the PDF).

A great example in support of this interpretation of induction is Aristotle’s comment on goodness (aretē), in the Eudemian Ethics. There, he states that it is “the best disposition or state or faculty of each class of things that have some use or work.” He then gives us reason to believe that he has identified the nature of goodness by using the example of a coat’s goodness, which belongs to a coat in virtue of it carrying out its particular function or use. The inductive generalization about goodness states the essence of the subject: what makes “goodness” the kind of thing it is; is open-ended in that it extends to all kinds of particulars that have uses (not simply the one coat or house that could be used to form this induction); points out the element of similarity which justifies the induction (things having particular functions or uses); and is related to sensory data and memories about things having various functions.

The second interpretation on Aristotelian induction has been supported by the majority of philosophers who have commented on Aristotle, including contemporaries such as philosopher of science John D. Norton. (Norton has originated a “material theory of induction,” and thinks that Aristotle supports the enumerative induction viewpoint (page 3 of the PDF).) It is supported by eight of the 12 uses of the term “induction” in the two-book Prior Analytics, specifically (PrA 2.23).

In this interpretation, induction is a kind of deduction that is validated by complete enumeration of cases. It cannot extend beyond the particulars used in forming the generalization, and if it does not include all of them then it is invalid. It is primarily a matter of deducing a property (major extreme) or feature to belong to all the particulars of a class (middle term) by arguing that the same property belongs to one or some of the class’s particulars (the minor extreme). This is then perfected by adding “etc.” or “and these are all the particulars.”

PrA 2.23 has baffled students of Aristotle and commentators alike, while at the same time it has been regarded as the chief chapter on Aristotelian induction. In the first paragraph, Aristotle states his often-repeated claim that there are two kinds of ways of having convictions: induction and deduction. In the very next sentence, the start of the second paragraph, he seems to contradict this by claiming that induction really is just a type of deduction.

McCaskey’s resolution of this dilemma not only clears Aristotle’s name from the list of enumerative inductivists, but also teaches us the process by which induction supplies the premises for deductions.

I’ll cover the approach of McCaskey’s resolution in Part 2.